Joint Perception and Control as Inference with an Object-based Implementation
This work addresses the challenge of integrating perception and decision-making in reinforcement learning for partially observable environments, offering a novel framework with potential applications in robotics and AI.
The paper tackles the problem of combining perception and control in partially observable environments by introducing a joint framework called Perception and Control as Inference (PCI), with an object-based instantiation (OPC) that achieves good perceptual grouping quality and outperforms baselines in accumulated rewards.
Existing model-based reinforcement learning methods often study perception modeling and decision making separately. We introduce joint Perception and Control as Inference (PCI), a general framework to combine perception and control for partially observable environments through Bayesian inference. Based on the fact that object-level inductive biases are critical in human perceptual learning and reasoning, we propose Object-based Perception Control (OPC), an instantiation of PCI which manages to facilitate control using automatic discovered object-based representations. We develop an unsupervised end-to-end solution and analyze the convergence of the perception model update. Experiments in a high-dimensional pixel environment demonstrate the learning effectiveness of our object-based perception control approach. Specifically, we show that OPC achieves good perceptual grouping quality and outperforms several strong baselines in accumulated rewards.